Older Adult Activity Recognition via Wearable Sensor Data

UCI Machine Learning Repository hosts this dataset on activity recognition for healthy older individuals, collected using a batteryless wearable sensor. The dataset contains time-series sensor data for classifying daily activities. The original source and collection date are not specified.

Use Cases

• Classify daily activities like walking, sitting, or standing using raw accelerometer and gyroscope time-series data.
• Train models to detect transitions between activities from sequential sensor readings.
• Analyze activity patterns and durations specific to an older adult demographic.
• Benchmark lightweight activity recognition algorithms designed for low-power, batteryless sensor systems.

Strengths

• Focus on a specific demographic (healthy older people), which is valuable for targeted research.
• Data collected using a batteryless wearable sensor, indicating a real-world, low-power application context.
• Hosted on the UCI Machine Learning Repository, a trusted source for benchmark datasets.

Limitations

• The exact number of participants, recording sessions, and total data points are unknown.
• Lack of detailed column descriptions and sample data hinders immediate analysis.
• Geographic and demographic specifics of the participant pool are not provided, which may limit generalizability.

Provenance

Source

UCI Machine Learning Repository

Collection Method

Data gathered using a batteryless wearable sensor worn by healthy older individuals.